Minutes of the 2nd FROST-2014 meeting
Moscow, 16-18 April 2012
VENUE: Hotel Marco Polo (http://www.presnja.ru/eng), Spiridonjevskij per., 9, Moscow, Russia
Day 1: 16 April
After a brief Welcome from Roshydromet by Igor Shumakov, Deputy Head of Roshydromet, the Agenda was adopted.
Introductory session
Chair: Philippe Steiner
Nanette Lomarda greeted the participants on behalf of the WMO and presented the talk “Welcome from WMO and WWRP information: FROST2014-related outcomes of the 5th meeting of the WWRP/JSC (Geneva, 11-13 April 2012)” (see http://frost2014.meteoinfo.ru/media/present/2meeting/FROST 2014_JSC5.pdf). Shortly before the meeting (on 12-13 April 2012) FROST-2014 was endorsed by the WWRP Joint Scientific Committee as a WWRP RDP/FDP.
Dmitry Kiktev outlined the progress made in the project during the past year in his presentation “Overview of FROST-2014 progress in 2011/2012: current disposition and preparations (observational network, available model output, archiving, visualization etc)” (http://frost2014.meteoinfo.ru/media/present/2meeting/FROST2012-Kiktev.pdf).
Discussion:
P.Joe noted that more solid precipitation gauges were needed for the project.
G. Isaac added that (fresh) snow density observations were needed as well.
Evgeny Vasilyev presented “Experience of meteorological forecasting for test sport events in Krasnaya Polyana in February-March 2012” (http://frost2014.meteoinfo.ru/media/present/2meeting/Vasilev-16.4.2012.pdf).
Discussion:
G. Isaac: How do you predict visibility?
It was answered that visibility was predicted subjectively.
P. Joe: That state (subjective visibility forecasting) could be a benchmark for the further project impact assessments.
Section “Observations – state of affairs and prospects with respect to FROST-2014”
Chair: Paul Joe
Alexander Melnichuk desribed “Radar facilities for Sochi-2014” (http://frost2014.meteoinfo.ru/media/present/2meeting/Radar facilities 2012.pdf).
Discussion:
B. Bica: What kind of radar data are available (Raw data? Format? Resolution? Quality)?
A. Melnichuk: Radar information will be available in autumn 2012. Both raw and BUFR data can be used. It is planned to start with BUFR.
P. Joe: Where will the second Micro-rain vertically pointing radar (MRR) be installed?[1]
A. Melnichuk: On a mobile platform. The exact location will be determined by the Chief Meteorologist V.Lukyanov.
G. Isaac commented that precipitation type (rain vs snow) need to be distinguished.
Arkadi Koldaev talked on “Vertical remote sounding of the atmosphere for Sochi-2014” (http://frost2014.meteoinfo.ru/media/present/2meeting/presentation FROST-2014_koldaev.pdf).
Discussion:
E.Kadygrov: The region of Sochi-2014 Olympics is a complex terrain. The orographically induced noise might distort the observations obtained from the mobile platform mentioned in the presentation. Do you use any acoustic protection screen for these observations?
Have you tried to compare your vertical sounding data with radiosonde observations?
The MTP-5 microwave sounder scans the atmosphere at different angles, so a really flat terrain is needed for this sensor, which is hardly available at Krasnaya Polyana and nearby.
A.Koldaev: No acoustic protection is necessary for the measurements because an open observation site with flat neighbourhood will be looked for for the mobile platform.
No comparison with radiosonde data has been made by now.
Nikolai Bocharnikov presented state of the art for a subset of automatic meteorological stations (AMS) near Sochi (http://frost2014.meteoinfo.ru/media/present/2meeting/Bocharnikov_FROST_16.04.2014.pdf ).
Discussion:
P. Steiner: Are the observation data quality controlled?
N. Bocharnikov: At the moment not, but the quality control can be implemented.
G. Isaac stressed the need for very high observation frequency (e.g. down to 1 sec for ski jumping).
N. Bocharnikov: There are 3-D wind sensors at sky jump. Now, meteorological parameters are measured with 10-minutes resolution, but they are transformed into temporal means, maxima or minima before their communication. From the hardware point of view, there are no limitations for much higher time resolution.
P. Nurmi interested if any road meteorological stations would be installed in the region.
N. Bocharnikov informed the participants that in April 2012 four road meteorological stations would be installed and several extra stations might be added later.
Dmitry Moisseev: “Recent advances in radar-based hydrometeor classification, what can we do during FROST-2014” (http://frost2014.meteoinfo.ru/media/present/2meeting/Present_Moisseev_FROSt.pdf ).
Discussion:
Several questions were asked:
G. Isaac: Can you distinguish the shape of falling particles?
N.Shakina: Do you have a software and results for freezing rain and drizzle?
D. Moisseev answered the questions positively, but noted that today this activity is at the research stage.
Peter Romanov presented an update on “NOAA snow satellite data analysis” (http://frost2014.meteoinfo.ru/media/present/2meeting/romanov_FROST-2014_April_16_2012.pdf).
Discussion:
M.Tsyrulnikov: Can you distinguish between 1-m and 2-m snow cover from space?
P. Ronanov: The snow cover is restored using brightness temperature. Most microwave algorithms can distinguish between situations with “snow < 1 m” and “snow >1 m”.
Alexander Shmelkin presented the «Meteofon» project (AMS installed on cell-phone network towers) (http://frost2014.meteoinfo.ru/media/present/2meeting/Меteofon_FROST.pdf).
Discussion:
M.Tsyrulnikov: Have you assessed the quality of your data?
A.Shmelkin: We transfer the observational data to Roshydromet and the conclusion of its quality should be given by Roshydromet
D. Kiktev: Can you improve the time resolution of your data? Can any additional sensors be installed at the towers?
A.Shmelkin: The resolution can be increased down to 1 min and finer. Additional sensors can be plugged in via the communication nodes «Weather Box». This is only question of money and extra-weight.
Paul Joe. Discussion (http://frost2014.meteoinfo.ru/media/present/2meeting/FROST14 Observations Discussion Points.pdf):
1) It was mentioned in the talk by E.Vasilyev that large differences were observed between two consecutive snow-cover depth observations. A possible reason for this is the application of acoustic sensors. Their performance can be affected by the presence of the falling snow.
A. Koldaev noted that ultra-sonic sensors’ beam can penetrate through fresh snow, so that sensors of another type should be used.
2) P. Steiner asked about the formats of available AMS observation data. D. Kiktev answered that XML-format is used and access has been provided to interested project participants.
3) It was noted that some sensors (Meteofon AMSs) are mounted at heights different from the usual 2m and 10m.
A. Shmelkin: The cell-phone network towers are of 25 to 75 m height. The sensors can be mounted at 3m-beams distantly from the towers to minimize the disturbances of the flow.
D. Kiktev: Non-standard heights for the sensors were used to account for shading effects of high trees and other hindrances around the towers. To some extent the additional elevation above the makes the observations more representative.
G.Isaac noted that the actual height of observations should be used in the models.
P.Joe: If we have 2m of snow cover, what is the observation height?
The participants agreed that all available observations can be used, and non-standard heights of observations is more a problem of models.
Section «Nowcasting activities – status and plans for FROST-2014»
Chairs: Paul Joe, George Isaac
George Isaac presented “SNOW-V10 experience in the context of FROST-2014: nowcasting and observations” (http://frost2014.meteoinfo.ru/media/present/2meeting/IsaacFrost2014Presentation1April2012.pdf), mentioning, in particular, that a special issue of PAGEOPH journal is being prepared, where 18 papers with SNOW-10 results are going to be published this year.
Discussion:
A.Montani mentioned the possible utility of poor man’s nowcasting ensembles. He was also interested if the gondola observations were used in operational forecasting activity.
G. Isaac answered that gondola information was not used for operation services.
Discussion about the forms of data presentation to weather forecasters:
- The forecasters prefer meteograms to 2D-fields.
- It is useful to supplement the current meteograms with previous meteograms from different models and present observation data at the same plot. Then forecasters will have a hint for an appropriate degree of confidence to the participating forecasting technologies.
Regarding the very high resolution models promoted by G. Isaac, D. Kiktev recalled that M. Mueller from the University of Basel demonstrated that after post-processing, e.g. for 2m-temperature, the difference in skill between low- and relatively high-resolution meso-scale models might become negligible (http://www.seevccc.rs/rnd/presentation-expertws/M_Muller_Predictability_of_scales_what_NWP_can_tell_us_for_climate_downscaling.pdf).
Benedikt Bica “INCA: state of affairs and prospects with respect to FROST-2014” (http://frost2014.meteoinfo.ru/media/present/2meeting/INCA_for_Sochi_2014 - short.pdf).
Discussion:
P. Joe: What are INCA requirements to the observational data quality?
B. Bica: INCA makes quality control of observations and removes bias. But radar data are needed to tune the system.
M. Tsyrulnikov: Spatial interpolation of precipitation fields gives rise to non-desirable smoothing. How does INCA cope with this?
B. Bica: Interpolations are performed only for gauges, producing smooth fields. The details are added by, first, introducing elevation-dependent corrections, and second, by using hi-res radar data.
Tatyana Bazlova presented “Nowcasting activity at the aviation centre of Adler” (http://frost2014.meteoinfo.ru/media/present/2meeting/Bazlova T_FROST_16.04.2012.pdf ).
Discussion:
N. Shakina asked several questions about techniques used in the presented system.
G. Isaac was interested if the system was verified.
Pertti Nurmi talked on “Potential road weather forecasting for Sochi-2014” (http://frost2014.meteoinfo.ru/media/present/2meeting/Road Weather_for_Sochi_2012-04-16_with-animation.pdf).
Discussion:
B. Bica: What kinds of meteorological observations are used in your system? How do you account for road composition and traffic density?
P. Nurmi: Road friction data and meteorological data along the entire road are necessary for training of the system. Later on, the system will use only the meteorological data from standard observation sensors. The road composition is accounted for via regression relations used in the system.
Paul Joe: Discussion
G. Isaac noted that from the point of view of verifications it is quite difficult to beat the persistence nowcasts.
D. Kiktev recalled some comments on his presentation at the 5th JSC Meeting. In particular it was noted that some kind of regulations on nowcasting/forecasting procedures for the Olympic meteorologists should be elaborated.
Day 2: 17 April
Section “Numerical Weather Prediction – status and plans for FROST-2014”
Chair: Stephane Belair
George Isaac discussed “SNOW-V10 experience in the context of FROST-2014: Numerical Weather Prediction” (http://frost2014.meteoinfo.ru/media/present/2meeting/IsaacFrost2014PresentationNWPApril2012.pdf)
Discussion:
- Meteograms – use of the nearest model grid point is not necessarily the best option. Elevation and configuration of relief should be taken into account.
- Verification results depend on the daytime.
- In SNOW-V10 INTW (INTegrated Weighted model using 1-km limited are model, 15-km regional GEM version and Observations) was nearly always the best.
D. Kiktev: Time resolution of nowcasting output is 10-15 min. What for the 1-min resolution data are needed?
G. Isaac: For verification. Meteorological values vary within a 10-min period. Various interval statistics (maxima, minima, means) are calculated on the basis of more frequent observations.
A. Montani noted that sometimes 15-km model was the best in the presentation results. This means that not only model resolution matters and, e.g. bad boundary layer scheme might be responsible for precipitation errors. He also mentioned role of model resolution for calculation of wind gusts.
Philippe Steiner presented the “COSMO contribution to FROST-2014” (http://frost2014.meteoinfo.ru/media/present/2meeting/20120417_sti_COSMO_FROST2012.pdf)
Discussion:
P. Romanov: Who are producers of SST and soil moisture data used in COSMO system?
P. Steiner: SST is taken from IFS, soil moisture adapted to surface temperature.
A.Montani presented “Ensemble forecasting for Sochi-2014 Olympics: the SOCHI-ensemble system from COSMO” (http://frost2014.meteoinfo.ru/media/present/2meeting/montani_frost_20120416.pdf).
Discussion:
G. Isaac commented that it would be of interest to compare results of 7-km COSMO ensemble with deterministic 2-km model version.
M. Tsyrulnikov: What forecasters like in your ensemble forecasts?
A. Montani:
a) Stability of results between the subsequent runs of the system. This year there was a heavy snowfall in Bologna. Deterministic models “hesitated” about the possibility of occurrence of this snowfall. EPS gave more stable prediction.
b) Aggregated information in the form of tables for various thresholds and subregions.
Inger-Lise Frogner presented the “HIRLAM contribution to FROST-2014” (http://frost2014.meteoinfo.ru/media/present/2meeting/Frogner_FROST_apr12.pdf).
Discussion:
M. Tsyrulnikov: You use the results of ECMWF EPS as boundary conditions. So, you have to wait for them. When is your forecast ready?
I.-L. Frogner: GLAMEPS starts at 6 and 18 UTC using the results of ECMWF EPS from 0 and 12 UTC correspondingly. The calculation takes about 6 hours.
M. Tsyrulnikov: Why does HIRLAM use two systems, GLAMEPS and HARMON EPS?
I.-L. Frogner: They have different resolutions. Eventually, HARMON EPS is planned to become the only HIRLAM EPS system.
D. Kiktev: Have you ever tried joint ensembles on the basis of HIRLAM and COSMO EPSs output?
I.-L. Frogner: It has not been tried.
Section “Numerical Weather Prediction – status and plans for FROST-2014” – continuation. Chair: Andrea Montani.
Jason Milbrandt presented “Environment Canada: status and plans for FROST-2014” (http://frost2014.meteoinfo.ru/media/present/2meeting/Milbrandt-FROST-Moscow_Apr2012.ppt).
Discussion:
P. Steiner: You have data assimilation (DA) only for the globe. How old are the data for LAM?
J. Milbrandt: 7-8 hours. Besides of the global DA we have regional assimilation of observations.
P. Steiner: When will the new version of GEM be ready?
J. Milbrandt:
Future of 15-km GEM: tests - this year, operational runs - from the beginning of 2013.
Gdaly Rivin talked on “Recent activities with COSMO-RU” (http://frost2014.meteoinfo.ru/media/present/2meeting/20120417_RivRozAll_Frost_END.pdf).
Discussion:
A. Kislov: Your results for the case study with bora phenomenon were shown for a quite large region. Are you sure that this was really bora?
G. Rivin: Yes, it can be seen both in 7-km and 2.2-km runs. Cross-sections of wind and temperature fields also support this conclusion.
Michael Tsyrulnikov talked on the “Development of the 3D-Var/COSMO based data assimilation system for the Sochi area” (http://frost2014.meteoinfo.ru/media/present/2meeting/tsy_FROST14_meeting2012.pdf ).
Discussion:
S. Belair: Assimilation of surface data is important for short-range forecasts. What is the situation with surface data assimilation?
M. Tsyrulnikov: SSTs are taken from GME, the snow - from NOAA.
D. Kiktev: Will you be able to assimilate the future radar radial winds data?
M. Tsyrulnikov: Yes, it will be possible.
S. Belair presented “Comments/Input of CAS/WWRP WG on Mesoscale Weather Forecasting Research” (http://frost2014.meteoinfo.ru/media/present/2meeting/20120417_BELAIR_COMMENTS.pdf).
Stephane Belair: Discussion.
P. Nurmi raised the issue of training in the use of forecast verification for Olympic forecasters.
I.-L. Frogner: We are working on a training course in ensemble forecasting and can provide access to it for the forecasters.
H. Astakhova: There should be a consistency between the forecast products used for training and real operations.
G. Isaac: It is important to know what is necessary for forecasters and what they must give to end-users. E.g., on one hand users don’t want probabilities, on the other hand they need visibility that is missed in ensembles.
A. Montani: To give a hint on how to present the information to forecasters one can look at our web-site with ensemble products. A simple form we started from was just to mention the probabilities writing under the meteogram words “low” or “high” probability.
S. Belair: Who will prepare graphics with forecast results - Roshydromet or the project participants?
D. Kiktev: For higher reliability and elaboration of efficient forms of forecast products it is preferable to supplement graphical products prepared by Roshydromet with forecast graphics prepared by the participants.
Day 3: 18 April
Section “Verification, transfer of technologies, impacts”.
Chairs: Nanette Lomarda and Pertti Nurmi
Paul Joe talked on “Transfer of technologies to forecasters: experience of Vancouver-2010” (http://frost2014.meteoinfo.ru/media/present/2meeting/JOE FROST14 Technology Transfer to Forecasters.pdf).
Nanette Lombarda presented “Input of CAS/WWRP WG on Societal and Economic Research” (http://frost2014.meteoinfo.ru/media/present/2meeting/WGSERA_FROST-2014.pdf).
Discussion:
D. Kiktev commented that relevant quantitative measures of societal and economic impacts are of great interest for the project and new ideas are needed.
Pertti Nurmi presented “Input of CAS/WWRP WG on Verification: Vision of verification setup for Sochi-2014” (http://frost2014.meteoinfo.ru/media/present/2meeting/FROST-14_Verification_Vision_2012-04-18.pdf).
Discussion:
G. Isaac noted that verification thresholds should be defined on the basis of end user (judges, event managers etc) requirements. Some thresholds might be unexpectedly changed.
D. Kiktev:
- A considerable part of the SNOW-V10 thresholds is thought to be applicable for FROST-2014.
- Conditional verification was not mentioned in the presentation.. Is it included in the verification set-up?
P. Nurmi: Conditional verification was not included.
G. Isaac: In SNOW-V10 there was no conditional verification either.
Nanette Lomarda and Pertti Nurmi: Discussion
G. Isaac raised the question on uncertainty in observations that can be to some extent accounted for via observational instrument intercomparison.
D. Kiktev noted that with respect to the uncertainty in observations not only observation errors matter, but temporal and spatial statistical structure of meteorological fields as well. High-resolution measurements can be useful for these assessments.
M. Tysrulnikov: These are so-called representativity errors.
Observation errors maybe site specific, field specific and event specific as well.
D. Kiktev mentioned that the basic verification package for FROST-2014 was still an open-ended question and that this issue was under examination. The considered options are MET (Model Evaluation Tool, developed by UCAR/NCAR) and VERSUS (developed by COSMO community).
Section “Project update and future steps (Discussion)”
Chair: Dmitry Kiktev
1) Project Goals
It was proposed by the chair to review the list of the FROST-2014 goals (presented in the project Concept Paper) and check if any changes were needed.
P. Joe suggested to add one more goal to the list: «To develop a comprehensive information resource of alpine winter weather observations». The participants agreed with this suggestion.
2) Project Milestones
Brief information on the project achievements in 2011 was presented by D. Kiktev:
- Observations and nowcasting (Working Group 1): In total, 48 stations with near-surface meteorological observations were deployed in the Sochi region. However, stability of the data flow and AMS data quality control are issues that must be urgently tackled. A Doppler radar on Akhun mountain is being installed (Data flow from the Doppler radar is to be available in autumn 2012). Additional radar data for Caucasian region (in particular, another new Doppler radar in Mineralnie Vody) are expected to be available in autumn 2012. During the 2011/2012 winter season the observation system was not well-established enough for objective nowcasting.
- Numerical Weather Prediction (Working Group 2): Tables with achievements of the project participants in 2011 compiled by M. Tsyrulnikov for the NWP-project component were presented (Annex 1).
- IT-support (Working Group 3): Observational and forecast/nowcast data storage with Internet-access for the participants was established for the project needs. Web-site of the project, http://frost2014.meteoinfo.ru with the first version of the project visualisation tools has been developed. Data flows from the AMSs of Sochi region and from real-time forecast technologies COSMO-7 (with and without data assimilation), COSMO-2, SOCHMEL EPS were established. - Products and training (Working Group 4): Training for Olympic forecasters is being carried out on the regular basis and participants are involved into provision of meteorological services for test sport events. Despite this, training is getting a crucial issue. Efficiency of the training is considered to be enhanced.
For the next FROST-2014 year (2012/2013) the following progress is expected to be made:
- Observations and nowcasting:
The participants of the nowcasting project component were asked to specify their input to the project Implementation Plan and to fill in the tables with nowcasting milestones (Annex 4).
- Numerical Weather Prediction: Tables with the milestones for 2012 compiled by M. Tsyrulnikov for the NWP-project component were presented (Annex 2,3). It was also noted that availability of TIGGE products to FROST-2014 was discussed with TIGGE project manager and the Chief of WMO World Weather Research Division during the recent 5th meeting of the WWRP Joint Scientific Committee. Some formal actions are needed from Roshydromet in this direction.
P. Nurmi informed the participants that FMI would contribute to the NWP-project component using deterministic version of HARMONIE system with 2.5km resolution and this could be added to the NWP-milestones table.
- IT-support: Further integration of data flows of observations (AMSs, radar data, profiler information) and forecast/nowcast data; unified graphical presentation of available observational and forecast data; provision of the project participants with access to the time series and the real-time observations and forecasts/nowcasts; update of web-based graphical tools for presentation of FROST-2014 products and their verifications. Web-forum facilities to be added to http://frost2014.meteoinfo.ru. - Products and training: Training on the use of FROST-2014 products will be organized during the next training course for Olympic forecasters (tentatively in October 2012). FROST-2014 products will be used during the next sport test events in winter 2012/2013.
Comments and Discussion:
P. Nurmi informed the participants that several courses on verification, including video and self-learning guidances are available.
P. Joe: Development of training plan (eMaterial, web video training, classroom training, golden cases for ski jump, downhill etc) is needed.
D. Kiktev noted that some kind of regulations on nowcasting/forecasting procedures for the Olympic meteorologists should be elaborated. He also identified several areas where participation of international lecturers at the training for the Olympic forecasters is desirable:
- interpretation of Doppler radar data;
- nowcasting methods;
- interpretation of ensemble forecasts.
D. Moisseev expressed his readiness to take part at the nearest training course (tentatively in October 2012) as a lecturer. Participation of other experts is to be confirmed.
3) Observations: available and needed
B. Bica: There is no use working with INCA while radar data are not available.
A. Melnichuk: Radar data will appear this summer, but they must be tested. So they will be ready for use in autumn 2012. A. Melnichuk also noted that max reflectivity map can be provided for INCA.
G. Isaac: An observational site with several gauges is worth arranging somewhere at altitude 2000 m, so that intercomparison of observations can be carried out.
D. Kiktev was not too optimistic about the flexibility of Roshydromet that was needed in this matter. He mentioned that probably the local Olympic Organizing Committee might help in support of this activity and in negotiations with sport venue owners.
G. Zaimskikh noted the need of closer coordination between the Olympic forecasting operations, FROST-2014 implementation and the Organizing Committee activities.
The participants noted that the already existing and planned observational set-up in the Sochi region in some respects should be enhanced, specifically:
- A precipitation occurrence sensor system (POSS) is considered to be important for more reliable monitoring of solid precipitation in Krasnaya Polyana;
- It is highly recommended to combine the already installed Micro-rain vertically pointing radar (MRR-2) with installation of a solid precipitation gauge at the same location;
- P. Joe mentioned that observations of snow density and types of precipitation particles (e.g. snow video imager) should be added. This information is needed to validate dual pol radar products;
- More reliable sensors of snow height are needed.
4) Quality control
D. Kiktev noted that at the moment the available AMS observations passed through only a very simplified quality control (QC). Enhancement of the QC was mentioned as one of the urgent tasks for the project organizers (as some meta-information about the stations today is available only within Roshydromet).
M. Tsyrulnikov: INCA makes its own quality control. Maybe it can use raw observation data?
B. Bica: INCA can make quality control but it takes time.
G. Isaac: Quality control was made in real time in Vancouver. The Canadian nowcasting system makes neither quality control nor obs bias correction - as observations are assumed to be calibrated.
5) Forecast data exchange
Unified longitude-latitude grids were suggested in the project Concept Paper for transfer of forecast fields between the FROST-2014 participants.
A.Montani and G. Isaac expressed their concern with respect to the unified grids for the forecast data exchange as a part of useful information might be lost due to the interpolation from original model grids to the unified grids.
D. Kiktev insisted that gridded data should be transferred on unified grids. The arguments for this were as follows:
- The primary object of verification in FROST-2014 will be forecast/nowcast time series
associated with station locations. For verification purposes the project participants will transmit their pointwise forecast time series associated with the AMS locations in XML-format. In this case the best way the gridded data are associated with station locations (via interpolation or using the nearest grid point) can be decided on manually by the producers of these forecasts. - Grid-point fields on the unified lat-lon grid will be used mostly for unified visualization of forecast fields at the project web-site.
- To deal with a variety of model grids would be technologically too resource-consuming.
Free access to all the project forecasts for all the participants was suggested. No objections were expressed.
6) Data representation
P. Joe: How is it planned to present the project data?
D. Kiktev: On the web. Some forecasters ask us for aggregation of various kinds of information into one tool with unified presentation of results from the engaged forecasting systems. We work on this. Observations, forecast maps and meteograms should be integrated into the same web-tool (http://frost2014.meteoinfo.ru). Aggregated “forecasters tables” should be prepared by the next winter test events.
P. Joe: Geodata for roads, towns, names, etc are needed.
7) AOB
M. Tsyrulnikov suggested the participants thinking about the project logo.
Closing of the Meeting
D. Kiktev spoke words of gratitude to the participants for their interest and involvement into the FROST-2014 project. He expressed the hope that the project would be interesting both for participants and organizers.
